Skip to main content
NCBI home page
Psychiatry, Psychology, and Law logoLink to Psychiatry, Psychology, and Law
. 2020 Nov 10;28(4):531–545. doi: 10.1080/13218719.2020.1805811

When punishment goals moderate and mediate the effect of clinical reports on the recidivism risk on prison sentences

Anta Niang a,b,c,✉, Chloé Leclerc d, Benoît Testé c
PMCID: PMC9090404  PMID: 35558147

Abstract

This research examined whether an individual’s endorsement of punishment goals moderates and mediates the effect of a clinical assessment of recidivism risk on the length of prison sentences. We measured participants’ endorsement of punishment goals, both before they read a criminal case (i.e. a priori endorsement), and after they had read it (i.e. case-specific endorsement). As expected, the effect of the clinical report’s conclusion on participants’ sentencing decisions was moderated by a priori endorsement of incapacitation as a punishment goal. Participants who expressed strong (versus weak) a priori endorsement of this punishment goal were influenced by the report’s conclusion on the risk of recidivism. In addition, when the clinical report concluded that the offender had a high risk of recidivating, participants expressed greater case-specific motivation to incapacitate him. Finally, the clinical report’s conclusion had an indirect effect on the severity of the sentence through case-specific endorsement of the incapacitation punishment goal.

Keywords: clinical report, incapacitation, jury decision making, punishment goals, risk of recidivism, sentence

Introduction

In France, as in numerous other countries, criminal cases frequently include evaluations of a defendant’s/offender’s psychological and psychiatric state (Bordel et al., 2004). The experts who make these evaluations are regularly appointed by the judge and paid by the State. This is not the case in adversarial proceedings, as practiced in common law countries (e.g. Canada, Spain, United Kingdom, United States) where experts are commissioned and paid by the parties, which may have a greater impact on the 'quality of the expertise and the neutrality of the expert' (Oytana, 2016, pp. 5–6).

An important aspect of many of these evaluations is to assess an offender’s risk of recidivism (Monohan & Steadman, 1994; Moulin & Palaric, 2013; Shuman & Sales, 1998), as preventing recidivism has been a major preoccupation in both Europe and North America for the last 30 years (Harcourt, 2011; Mary, 2001). Given this focus on recidivism, several studies have investigated possible links between recidivism risk assessments and decisions relating to offenders. However, none of these studies allows clear links to be drawn between the sentences recommended by jurors and the punishment goals their decisions are intended to achieve. Nor do they show whether a report about an offender’s risk of recidivism influences all jurors in the same way, or whether this influence is stronger on some jurors than on others. Hence, the present study investigated whether the influence of clinical assessments of recidivism risk depends on the punishment goals a juror endorses, in particular the importance given to incapacitating the offender.

Interest in assessing recidivism risk grew substantially with the emergence of what has become known as the ‘New Penology’ (Feeley & Simon, 1992) and the growing focus on preventing recidivism. One result of this has been the creation of numerous tools for evaluating the risk of recidivism (Desmarais et al., 2016; Viljoen et al., 2018). In this context of actuarial justice, the aim of assessing the risk of an offender recidivating is to try and identify and control individuals assessed as being at a high risk of reoffending (Van Wingerden et al., 2014). One criticism of this approach is that it could lead to harsher punishments for these individuals. Indeed, research into the effects of providing jurors with information about an offender’s recidivism risk suggests that recidivism risk assessments affect the severity of their sentencing decisions. Studies in the USA of capital offences show that expert reports indicating a high recidivism risk increase perceptions of the offender’s future dangerousness, and that this effect is particularly strong in the case of clinical reports (compared with actuarial reports; Krauss & Lee, 2003; Krauss et al., 2004; Krauss & Sales, 2001). Other studies have shown that mock jurors are more likely to recommend psychiatric incarceration when they are presented with an expert report indicating a high risk of the offender recidivating (Krauss et al., 2012). Krauss and Scurich (2014) showed that reading an expert report indicating a moderate risk of recidivism was enough to increase decisions in favor of psychiatric incarceration, compared with seeing a report indicating a low recidivism risk. Information about recidivism risk also affects decisions regarding civil commitment,1 or post-sentence detention with participants being more likely to recommend civil commitment (Carlsmith et al., 2007) or post-sentence detention (Tostain & Lebreuilly, 2013) for offenders assessed as having a high recidivism risk.

Curiously, the influence of recidivism risk assessments on the length of prison sentences has not been studied, even though this is the most common type of sentence given to perpetrators of serious offenses. Moreover, to date, there has been no research into whether individual variables can moderate the influence of recidivism risk assessments on jurors’ decisions. The present study helps remedy this situation by examining jurors’ endorsement of different punishment goals as potential moderating variables of the effect of an offender’s risk of recidivism on the severity of sentencing decisions.

The notion of punishment goals raises the question of ‘why punish?’ (Cusson, 1987). From a philosophical point of view, the issue can be viewed in terms of either retributivism (Kant, 1781/1980) or utilitarianism (Beccaria, 1764; Bentham, 1802). Retributivism is founded on the ancient law of talion and the idea that punishment should be proportional to the crime (i.e. just deserts, Carlsmith, 2008), whereas utilitarianism (also known as consequentialism) aims to impose a ‘useful’ punishment that is beneficial for society and/or the offender’s social reintegration (Tostain & Lebreuilly, 2013). On an individual level, jurors may associate the sentence imposed with a desire to restore the moral equilibrium that has been disturbed by a crime (retribution goal), but they may also be motivated by the utilitarian ideals of rehabilitating (rehabilitation goal), dissuading (deterrence goal) or neutralizing (incapacitation goal) the offender (Atanasova-Denié & Tostain, 2008; Carlsmith, 2008; Carlsmith & Darley, 2008; Carlsmith et al., 2007; Tostain, 2007).

Punishment goals are an important factor when determining sentences (Carlsmith et al., 2002; Darley et al., 2000; Hamilton & Rytina, 1980; Jacoby & Cullen, 1998), with the importance given to different punishment goals varying between individuals and according to the nature of the crime and the characteristics of the offender (Orth, 2003; Vidmar & Miller, 1980) or of the case (Warner, Davis, Spiranovic, Cockburn, & Freiberg, 2019). In psychological research, punishment goals have been operationalized both as dependent variables and as independent variables. Punishment goals have been manipulated through instructions given to participants (Darley et al., 2000; McFatter, 1978) or the characteristics of the case (Carlsmith, 2008; Carlsmith et al., 2002; Darley et al., 2000). Other studies have measured individuals’ endorsement of punishment goals via items completed by participants either before (Graham et al., 1997; Orth, 2003) or after reading a criminal case (Atanasova-Denié & Tostain, 2008; Carlsmith et al., 2002; Oswald et al., 2002; Tostain, 2007; Warner et al., 2019). The present study examined whether endorsement of punishment goals moderates (i.e. a priori endorsement of punishment goals) and mediates (i.e. case-specific endorsement of punishment goals) the effect of a clinical recidivism risk assessment on the severity of mock jurors’ decisions on prison sentences for an offender.

Present study

The present study was conducted in France, where jurors participate directly in sentencing decisions, as they do in many other countries, including Belgium and Italy. In France, the court of assizes is made up of three judges and a popular jury of six citizens drawn by lots. The jurors sit alongside the judges during the trial and are exposed to both factual and testimonial evidence. At the end of the trial, they deliberate with the judges on the guilt of the accused and the sentence to be imposed. Throughout the deliberation, the jury must keep in mind the principle of intimate conviction found in several continental systems (Dumas & Esnard, 2019; Esnard & Dumas, 2019; Esnard et al., 2013). With regard more specifically to sentencing, a majority must be obtained at the end of the deliberations, thus demonstrating not only the complementary role of jurors (Arce et al., 1996) but also their importance in sentencing. This system contrasts with the criminal procedure in Common Law where the judge makes sentencing decisions.

Of all the different punishment goals, incapacitation would appear to be the best suited for examining the influence of recidivism risk assessments, as the aim of incapacitation is to prevent an offender committing further offenses (Carlsmith & Darley, 2008). Consequently, endorsement of incapacitation as a punishment goal should affect the influence a recidivism risk assessment has on a juror’s sentencing decisions. In addition, it is possible that not all jurors will be influenced by recidivism risk assessments to the same extent, as this influence should depend on the strength with which each juror endorses the goal of incapacitating the offender.

In the current study, we presented participants with a case of manslaughter in which a psychologist and a psychiatrist had conducted a clinical evaluation of the risk of the offender recidivating. We created two experimental conditions by manipulating the conclusion of the experts’ report in order to present the offender as having either a low or a high risk of recidivating (between-participants manipulation). Before presenting participants with the clinical report’s conclusion, we asked them to state the sentence they would impose on the offender. We then asked them to read the report’s conclusion and once again indicate the sentence they would impose. This methodology, involving repeated measures of the clinical report’s influence, was similar to the method used by Krauss et al. (2001, 2003, 2004).

Similarly, we measured participants’ endorsement of punishment goals on two occasions: before they had read the criminal case (i.e. a priori endorsement of punishment goals) and after they had read the case and the clinical report’s conclusion (i.e. case-specific endorsement of punishment goals). This gave us two measures of participants’ endorsement of punishment goals: a distal measure indicating their general, ideological endorsement of each punishment goal and a proximal measure indicating their case-specific endorsement of each goal (for similar methodologies, see Graham et al., 1997, and Weiner et al., 1997). Although this issue has not really been addressed in the literature, there are likely to be differences between a person’s a priori endorsement of punishment goals and the goals that person endorses in a specific case (Carlsmith, 2008; Garapon, 1997; Graham et al., 1997). Our study examined whether individuals’ punishment goals can both moderate and mediate the effect of a clinical assessment of recidivism risk on sentence length. We postulated that the influence of the clinical report on sentence length would be moderated by jurors’ a priori endorsement of punishment goals, and that the information about the offender contained in the report would affect their case-specific endorsement of punishment goals. Consequently, we expected case-specific endorsement of punishment goals to mediate the clinical report’s effect on sentence length.

We examined four hypotheses:

  • Participants will impose longer sentences when a clinical assessment of recidivism risk concludes there is a high (vs. low) risk of the offender recidivating (Hypothesis 1).

  • Participants’ sensitivity to the clinical assessment of recidivism risk will depend on their a priori endorsement of incapacitation as a punishment goal. More specifically, we expected the clinical report to have a greater impact on sentence length for participants whose a priori endorsement of incapacitation is strong (vs. weak; Hypothesis 2, moderation hypothesis).

  • The expert report’s conclusion will affect participants’ case-specific endorsement of punishment goals, leading to stronger case-specific endorsement of incapacitation in the high recidivism risk condition than in the low recidivism risk condition (Hypothesis 3).

  • The expert report’s conclusion will have an indirect effect on sentence length through case-specific endorsement of the incapacitation punishment goal (Hypothesis 4, mediation hypothesis).

Method

Participants

We collected data for our study via an online survey that was approved by a Research Ethics Board. All participants signed a consent form before completing the survey. Participants were 102 French citizens (70% women), aged between 18 and 58 years (M = 25), including 62 students from psychology and the others from various fields of study (psychology, law, ergonomics, history, neuroscience and sport).

Sensitivity power analyses performed with GPower (Faul et al., 2007) indicated that our sample could detect a small to medium effect size (i.e. f2 = .08) with 80% power and alpha of .05 in linear models including one tested predictor in a total number of four predictors (the largest number of predictors used in our analyses).

Procedure and design

The survey began by asking participants to indicate the extent to which four general punishment goals were consistent with their conception of a sentence. We then asked them to read a summary of a criminal case involving a 38-year-old man, Mr. Durand, who was on trial for killing his wife. Participants learned that Mr. Durand had stabbed his wife twice in a fit of jealous rage after she had told him that she had met another man. When he realized what he had done, Mr. Durand tried to save his wife’s life and called the police. Mrs. Durand was taken to hospital, but she died from her injuries. After reading the case, participants indicated the sentence they felt Mr. Durand should receive (initial sentence).

Participants were then presented with the conclusions of a psychological and psychiatric evaluation of Mr. Durand, requested by the judge, which we manipulated between participants. In the low recidivism risk condition, the experts concluded: ‘Despite the impulsiveness of his act, Mr. Durand has no difficulty controlling his feelings’ and ‘has a low risk of recidivism.’ In the high recidivism risk condition, the experts concluded: ‘Because of his impulsive nature, Mr. Durand finds it difficult to control his feelings’ and ‘has a high risk of recidivism.’ After reading the conclusion of the clinical report, participants rated items relating to their case-specific endorsement of punishment goals and again stated the sentence they felt Mr. Durand should receive (final sentence).2

Measures

A priori endorsement of punishment goals

Before reading the case, participants indicated the extent to which different general punishment goals (retribution, rehabilitation, deterrence and incapacitation of the offender)3 were consistent with their conception of sentences (e.g. retribution, ‘In my opinion, the general goal of a sentence is to make the offender pay for the crime he committed’; from 0 = ‘not at all consistent’ to 10 ‘fully consistent’). We used two items to assess each goal (see the Supplemental data section for more details) and combined the scores for each pair of items in order to obtain a score for a priori endorsement of each of the four punishment goals: retribution (r = .35, p ˂ .001), rehabilitation (r = .59, p ˂ .001), deterrence (r = .31, p ˂ .001) and incapacitation (r = .65, p = .002).

Case-specific punishment goals

After reading the case and the conclusions of the clinical report, participants indicated their case-specific endorsement of punishment goals. We used the same items as those for our measures of a priori endorsement of punishment goals but we put them into the context of the crime (e.g. retribution, ‘In the case of Mr. Durand, my preferred sentence would mean he pays for his crime’; from 0 = ‘not at all consistent’ to 10 = ‘fully consistent’). Again, we combined the scores for each pair of items in order to obtain a score for case-specific endorsement of each punishment goal: retribution (r = .54, p ˂ .001), rehabilitation (r = .49, p ˂ .001), deterrence (r = .51, p ˂ .001) and incapacitation (r = .82, p ˂ .001).

Sentence

Participants stated their preferred sentence for the offender on two occasions: immediately after reading the case (initial sentence) and after reading the conclusion of the clinical recidivism risk assessment (final sentence). Participants were told they could recommend a sentence of between 0 and 30 years’ imprisonment, as the maximum sentence for this type of case in France is 30 years. Participants were also told that the prosecutor had requested a sentence of 15 years’ imprisonment,4 but the offender’s attorney had asked for a more clement sentence. We asked participants to indicate both their preferred sentence and the minimum and maximum sentences they would find acceptable. Given the heterogeneity of the proposed sentences, we carried out a 90% winsorization (see Reifman & Keyton, 2010, for a definition and Monnery, 2016 (p. 85) for a winsorizing applied on sentence reductions) in order to limit the influence of extreme values. Hence, we replaced values below the 5th percentile and above the 95th percentile by the threshold value.

Results

Preliminary analyses

Preliminary analyses showed that participant’s age, β = −0.05, t = −1.19, p = .238, gender,

F = 0.008, p = .928, status, F = 1.33, p = .268, and field of study, F = 1.84, p = .165, did not impact the preferred sentence.

In addition, in order to ensure that the 15-year sentence recommended by the prosecutor does not lead to an entrenchment bias among participants, sentencing frequencies under each of the two conditions were examined. Results showed that most of the participants did not follow this recommendation (see Table 1 for more details).

Table 1.

Frequency of mock jurors sentences that follow the prosecutor's recommendation (%).

Sentence Low risk of recidivism
(n = 48)		 High risk of recidivism
(n = 54)
Initial    
 minimum 6.3 14.8
 preferred 27.1 14.8
 maximum 29.2 31.5
Final    
 minimum 6.3 14.8
 preferred 29.2 11.1
 maximum 27.1 27.8
Open in a new tab

Then, we began by examining whether a priori endorsement of punishment goals predicted participants’ initial preferred sentences (i.e. before they had read the clinical report’s conclusion). All the correlations between endorsement of retribution, deterrence and incapacitation were positive and significant (from r = .31 to r = .45, ps < .01). Rehabilitation was not related to the other goals (from r = −.14 to r = .08, ps > .14). A multiple regression analysis including all four general punishment goals as independent variables showed that endorsement of retribution, rehabilitation and incapacitation predicted the initial sentence (β = 0.20, t = 1.97, p = .051; β = −0.19, t = −2.05, p = .043; β = 0.22, t = 2.18, p = .032, respectively). Endorsement of retribution and incapacitation led to longer sentences, whereas endorsement of rehabilitation led to shorter sentences. Endorsement of deterrence had no effect on sentence length, β = 0.13, t = 1.30, p = .195.

We then examined whether participants’ case-specific endorsement of punishment goals predicted their final preferred sentence. All the correlations between case-specific endorsement of punishment goals (including rehabilitation) were positive and significant (from r = .56 to r = .71, ps < .001). Again, a multiple regression analysis showed that endorsement of retribution, rehabilitation and incapacitation predicted the final sentence (β = 0.50, t = 3.83, p < .001; β = −0.32, t = −2.62, p = .010; β = 0.23, t = 2.07, p = .041, respectively), but the effect of endorsement of deterrence was not significant, β = 0.19, t = 1.65, p = .102.

Effects of the clinical report and a priori endorsement of incapacitation goal on the final sentence

We examined Hypotheses 1 and 2 by using Process Model 1 to conduct a moderation analysis (Hayes, 2013) with final sentence as the dependent variable, clinical report conclusion as the independent variable (low recidivism risk = −1, high recidivism risk = +1) and a priori endorsement of incapacitation as the moderator variable (see Table 2 for a summary of the results). Despite some reluctance to use the pre- and post-test research design (Howard, 1980; Marsden & Torgerson, 2012; Robinson & Doueck, 1994), we have chosen to carry out our research in line with the design used by many studies on the impact of type of report (e.g. Krauss & Lee, 2003; Krauss et al., 2004; Krauss & Sales, 2001). So we entered initial sentence as a covariate in this model in order to control for the effect of the decisions participants made before reading the clinical report.

Table 2.

Effect of the clinical report and a priori endorsement of incapacitation on sentences.

  Sentence
Final minimal
(β)		 Preferred
(β)		 Final maximal
(β)
Clinical report 0.34* 0.33 0.26
Incapacitation 0.14 −0.05 0.06
Clinical report × Incapacitation 0.05 0.23* 0.29*
Open in a new tab

*p < .05.

Analysis of the participants’ minimum acceptable sentences revealed a significant main effect of clinical report, β = 0.34, t = 2.05, p = .043, 95% confidence interval, CI [0.01, 0.68], with participants in the high recidivism risk condition choosing a longer minimum sentence (Madjusted = 8.47, SE = 0.23) than participants in the low recidivism risk condition (Madjusted = 7.75, SE = 0.25). Neither the main effect of a priori endorsement of incapacitation, β = 0.14, t = 1.86, p = .066, CI [−0.01, 0.28], nor the interaction between the two variables, β = 0.05, t = 0.73, p = .468, CI [−0.09, 0.19] was significant.

Analysis of the participants’ maximum acceptable sentences showed that neither clinical report conclusion, β = 0.26, t = 0.75, p = .456, CI [−0.43, 0.95], nor a priori endorsement of incapacitation, β = 0.06, t = 0.39, p = .693, CI [−0.24, 0.36], had a significant main effect. However, the interaction between clinical report conclusion and a priori endorsement of incapacitation was significant, β = 0.29, t = 2.04, p = .044, CI [0.01, 0.58]. The conditional effects showed that participants who strongly endorsed incapacitation (+1 SD) chose longer maximum sentences in the high recidivism risk condition than in the low recidivism risk condition, β = 0.97, t = 1.97, p = .051, 95% CI [0.00; 1.96]. Participants who moderately or weakly (−1 SD) endorsed incapacitation were not influenced by the clinical report (β = 0.26, t = 0.75, p = .456, 95% CI [−0.42, 0.95]; β = −0.46, t = −0.93, p = .355, 95% CI [−1.43, 0.52], respectively).

Finally, analysis of the final preferred sentences showed that the main effects of clinical report, β = 0.33, t = 1.22, p = .227, 95% CI [−0.21, 0.87], and a priori endorsement of incapacitation, β = −0.05, t = −0.39, p = .692, 95% CI [−0.28, 0.19], were not significant. However, we again found a significant interaction between the two variables, β = 0.23, t = 2.06, p = .042, 95% CI [0.01, 0.45]. The conditional effects showed that participants who strongly endorsed incapacitation (+1 SD) preferred longer sentences in the high recidivism risk condition than in the low recidivism risk condition, β = 0.89, t = 2.32, p = .022, 95% CI [0.13, 1.66]. The effect of clinical report was not significant for participants who moderately or weakly (−1 SD) endorsed incapacitation (β = 0.33, t = 1.22, p = .227, 95% CI [−0.21, 0.87]; β = −0.24, t = −0.61, p = .542, 95% CI [−0.99, 0.53], respectively).5

Effect of the clinical report on case-specific punishment goals

We used univariate analyses of variance (ANOVAs) to examine whether the clinical report’s conclusion influenced participants’ case-specific endorsement of punishment goals (Hypothesis 3). Results (Table 3) show that the report influenced case-specific endorsement of both incapacitation and rehabilitation. When the clinical report concluded there was a high recidivism risk, participants were more likely to use the sentence as a way of protecting society from the offender and to view it in terms of helping rehabilitate the offender.

Table 3.

Effect of clinical report conclusion (high vs. low recidivism risk) on participants’ case-specific endorsement of punishment goals.

  High risk Low risk F(1, 100) p ŋp2
Retribution 6.06 (2.70) 5.81 (2.54) 0.23 .630 .00
Deterrence 5.56 (2.54) 4.83 (2.36) 2.25 .137 .02
Incapacitation 5.61 (3.09) 3.10 (2.81) 18.17 <.001 .15
Rehabilitation 6.42 (2.20) 5.13 (2.37) 8.24 .005 .08
Open in a new tab

Indirect effect of the clinical report on the final sentence through case-specific endorsement of punishment goals

The clinical report’s conclusion affected case-specific endorsement of two punishment goals: rehabilitation and incapacitation. Consequently, we examined Hypothesis 4 by using Process Model 4 (Hayes, 2013) to carry out multiple mediation analyses with endorsement of rehabilitation and endorsement of incapacitation as potential mediators, clinical report conclusion as an independent variable, final sentence as a dependent variable and initial sentence as a covariate.

The direct effect of the report on final minimum sentence was not significant, β = 0.09, t = 0.52, p = .606, 95% CI [−0.25, 0.44]. The effect of case-specific endorsement of incapacitation was significant, β = 0.18, t = 2.67, p = .009, 95% CI [0.05, 0.32], but the effect of case-specific endorsement of rehabilitation was not, β = 0.07, t = 0.86, p = .392, 95% CI [−0.09, 0.24]. Additionally, the report had a significant impact on both the case-specific endorsement of incapacitation, β = 1.22, t = 4.46, p < .001, 95% CI [0.68; 1.76], and the case-specific endorsement of the rehabilitation, β = 0.63, t = 2.87, p = .005, 95% CI [0.20; 1.07]. The indirect effect of clinical report through case-specific endorsement of incapacitation was significant, β = 0.23, 95% CI [0.08, 0.44], but its indirect effect through case-specific endorsement of rehabilitation was not significant, β = 0.05, 95% CI [−0.02, 0.21] (see Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Effect of clinical report on final minimal sentence through case-specific endorsement of incapacitation and rehabilitation.

*p < .05. **p < .01. ***p < .001.

The direct effect of the report on final maximum sentence was not significant, β = −0.17, t = −0.45, p = .653, 95% CI [−0.91, 0.57]. Again, the effect of case-specific endorsement of incapacitation was significant, β = 0.33, t = 2.28, p = .025, 95% CI [0.04, 0.62], but the effect of case-specific endorsement of rehabilitation was not, β = 0.03, t = 0.18, p = .861, 95% CI [−0.33, 0.39]. Additionally, the report had a significant impact on both the case-specific endorsement of incapacitation, β = 1.24, t = 4.44, p < .001, 95% CI [0.69, 1.80], and the case-specific endorsement of the rehabilitation, β = 0.64, t = 2.89, p = .005, 95% CI [0.20, 1.09]. The indirect effect of clinical report through case-specific endorsement of incapacitation was significant, β = 0.41, 95% CI [0.11, 0.85], but its indirect effect through case-specific endorsement of rehabilitation was not, β = 0.02, 95% CI [−0.15, 0.26] (see Figure 2).

Figure 2.

Figure 2.

Open in a new tab

Effect of clinical report on final maximal sentence through case-specific endorsement of incapacitation and rehabilitation.

*p < .05. **p < .01. ***p < .001.

Finally, the analysis of preferred sentences showed the same pattern, with a nonsignificant direct effect of clinical report, β = −0.03, t = −0.09, p = .921, 95% CI [−0.61, 0.55], a significant effect of case-specific endorsement of incapacitation, β = 0.32, t = 2.81, p = .006, 95% CI [0.09, 0.54], and a nonsignificant effect of case-specific endorsement of rehabilitation, β = −0.09, t = −0.64, p = .521, 95% CI [−0.37, 0.19]. Additionally, the report had a significant impact on both the case-specific endorsement of incapacitation, β = 1.28, t = 4.68, p < .001, 95% CI [0.74, 1.83], and the case-specific endorsement of the rehabilitation, β = 0.67, t = 3.04, p = .003, 95% CI [0.23, 1.10]. Again, the indirect effect of clinical report through case-specific endorsement of incapacitation was significant, β = 0.40, 95% CI [0.09, 0.81], but its indirect effect through case-specific endorsement of rehabilitation was not, β = −0.06, 95% CI [−0.22, 0.05] (see Figure 3).

Figure 3.

Figure 3.

Open in a new tab

Effect of clinical report on preferred sentence through case-specific endorsement of incapacitation and rehabilitation.

**p < .01. ***p < .001.

These results suggest that a clinical report concluding a high risk of the offender recidivating increased participants’ motivation to ensure the punishment protected society (i.e. case-specific incapacitation goal), which, in turn, affected the lengths of the minimum and maximum sentences they deemed acceptable, and their preferred prison sentences.

Discussion

The present study examined whether an individual’s endorsement of punishment goals moderates and mediates the effect of a clinical recidivism risk assessment on the length of sentence given to an offender. Results showed that the clinical report’s conclusion affected the minimum sentence that participants deemed acceptable, but we did not obtain a main effect of clinical report on the participants’ preferred sentences or maximum acceptable sentences. This result contrasts with earlier studies in which providing mock jurors with information indicating a high recidivism risk led to more severe decisions relating to offenders (Carlsmith et al., 2007; Krauss & Lee, 2003; Krauss et al., 2012; Krauss & Sales, 2001; Van Wingerden et al., 2014).6 However, these studies considered decisions that were directly linked to the ‘danger’ an individual might pose to society, and showed that a high recidivism risk led to more severe decisions regarding civil commitment or post-sentence detention (Carlsmith et al., 2007; Krauss et al., 2012; Krauss & Sales, 2001; Tostain & Lebreuilly, 2013).

To the best of our knowledge, our study is the first ever investigation of the impact of a clinical assessment of recidivism risk on prison sentence length. Results suggest that recidivism risk assessments have less influence on jurors’ decisions about prison sentences than on other decisions more intrinsically linked to the risk of an offender recidivating, such as recommending civil commitment or post-sentence detention. This finding contributes to the debates surrounding the effects of ‘actuarial justice’ and the ‘New Penology’ (Feeley & Simon, 1992), which give a central role to recidivism risk assessments. In the present study, the clinical report’s conclusions impacted the minimum sentence the participants deemed acceptable, suggesting that they took this information into account, probably applying the principle of precaution in order to protect society. Nevertheless, our results suggest that information about the risk of recidivism does not always have as great an effect as has often been supposed, at least in terms of the length of prison sentences. Further studies are needed to determine the conditions in which a clinical appraisal of recidivism risk influences the sentence given to an offender and to examine whether this process depends on the type of case being assessed.

Although the clinical report’s conclusions did not have a significant main effect on preferred sentences, some participants were influenced by the report. This is a major finding, as participants who reported strong a priori endorsement of incapacitation as a punishment goal felt that the offender should be punished with a longer prison sentence when the clinical report concluded there was a high (vs. low) risk of recidivism. Conversely, participants whose a priori endorsement of incapacitation was moderate or weak were not influenced by the clinical report’s conclusion. Hence, and as expected, the recidivism risk assessment did not influence all participants in the same way.

An unexpected result was that a priori endorsement of retribution as a punishment goal also moderated the effect of the clinical report’s conclusion on the maximum sentence that participants deemed acceptable. We did not obtain this moderation effect with respect to the minimum sentence or preferred sentence. Thus, it would appear that recidivism risk assessments not only influence jurors who strongly endorse incapacitation as a punishment goal, but may also influence jurors who strongly endorse retribution. Although this result may appear surprising, it can probably be explained by the fact that, like endorsement of incapacitation, endorsement of retribution as a punishment goal is related to a more punitive attitude toward criminals (e.g. Oswald et al., 2002).

This result also suggests that a clinical report on the risk of recidivism leads to punitive attitudes only with some of the jurors, which highlights the need to more accurately identify the individual variables that moderate the influence of recidivism risk assessments. Cognitive processes involved in the selection of information are a relevant source of explanation to better understand how individuals select the information presented to them and why the same information may be treated differently. Some researchers (Esnard et al., 2013; Esnard & Dumas, 2019) have shown, for example, that the principle of intimate conviction can lead jurors to confirmatory bias in decision-making. Thus, some jurors would have to select information consistent with their decisions. As pointed out by Dumas and Esnard (2019, p. 247), ‘individuals engaged in a motivational process are motivated to expose themselves to decision-consistent information and/or avoid decision-inconsistent information, in order to reduce the discomfort of dissonance arousal.’ Other studies (e.g. Epstein, 1994; Krauss et al., 2004; Krauss et al., 2012) have focused more on the ‘cognitive-experimental self-theory’ (Epstein, 1994), which refers to cognitive processes involving two types of reasoning (i.e. type of reasoning employed by individuals: rational versus experiential). But to date, only two of these studies (Krauss et al., 2004; Krauss et al., 2012) dealt with the impact of this type of expert report on legal decisions. These studies found that individuals who apply rational reasoning do not take into account the same expert information as those who apply experiential reasoning. However, in contrast to our study, they focused on the effects of different types of expert report (clinical vs. actuarial) and not on the report’s conclusion. Our study provides new data that help identify the individual determinants of the influence of clinical reports on legal decisions.

Finally, our study shows the need to consider punishment goals on both the abstract (a priori) level and the concrete (case-specific) level (Weiner et al., 1997). Weiner et al. (1997) reported identical and complementary impacts of two case-specific conceptions of punishment (retributivism and utilitarianism) on the severity of a sentence. By measuring endorsement of four case-specific punishment goals (deterrence, incapacitation, rehabilitation, retribution), we were able to show that the clinical report’s conclusion influenced two case-specific punishment goals (incapacitation and rehabilitation). This finding supports the theoretical view that case-specific punishment goals are highly malleable and context dependent. In addition, mediation analyses showed that only the case-specific incapacitation goal had a direct effect on length of sentence, and, as expected, the effect of the clinical report’s conclusion had an indirect effect on sentence length through this case-specific punishment goal. This indirect effect impacted both the participants’ preferred sentences and their minimum and the maximum acceptable sentences. Hence, the effect on sentence length of the report concluding a high recidivism risk appears to have been the result of the report increasing participants’ motivation to protect society by incapacitating the offender.

According to Weiner et al. (1997), ‘individual decisions regarding punishment goals tend to be situation specific’ and should therefore be measured in the context of the crime. This point of view calls into question the utility of measuring a priori punishment goals. However, our decision to measure punishment goals on an abstract and general level, as well as on a concrete and specific level, helped clarify the impact of the clinical report on the jurors’ sentences. In fact, our results show that punishment goals can both moderate and mediate the effect of a clinical report’s conclusion on a prison sentence, with a priori punishment goals moderating the effect and case-specific punishment goals mediating it. Thus, measuring both a priori and case-specific punishment goals is important in order to clarify the processes through which information about the risk of recidivism influences some individuals’ sentencing decisions. The higher an offender’s risk of recidivism, the more he or she is seen as a danger to society and the more individuals manifest a desire to protect society (Tostain & Lebreuilly, 2013).

A limitation of our study is that it was based on a single vignette and therefore one type of criminal case (domestic manslaughter). Future studies should determine whether comparable results are obtained with other types of case. In addition, our study population was not fully representative of the diversity of the population of jurors. Future studies should try to remedy this limitation. Nevertheless, we believe that our study provides important insights, not only for scholars working on the determinants of sentencing decisions, but also for psychologists and psychiatrists, as well as for judges and jurors, who have to make sentencing decisions in real cases every day.

Ethical standards

Declaration of conflicts of interest

Anta Niang has declared no conflicts of interest

Chloé Leclerc has declared no conflicts of interest

Benoît Testé has declared no conflicts of interest

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the Arts and Science Research Ethics Board at the University of Montreal and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Footnotes

1

Civil commitment is a process in which the judge can decide to send offenders who are considered to be dangerous to society to a psychiatric hospital.

2

Participants also rated other items that are not relevant to the present paper. The authors will supply these items upon request.

3

In this study, we did not study the impact of the combination of the different goals of punishment on the sentence. Although this issue has already received attention (see, for example, Niang et al., 2019), we believe that this issue should be further explored in future research.

4

It should be pointed out that during the French criminal trial, the prosecutor, who represents the interests of society, is required to recommend a sentence. However, jurors remain subject to the oath they have taken and so 'to decide according to the charges and the means of defence, according to [their] conscience and [their] intimate conviction with the impartiality and firmness required for a worthy and free man' (Article 304 of the French Code of Criminal Procedure).

5

We also examined whether other a priori punishment goals moderated the effect of the clinical report on the sentence. None of the interactions between the report’s conclusions and the other a priori punishment goals (retribution, rehabilitation and deterrence) was significant for either the minimum sentence (ps > .280) or the preferred sentence (ps > .074). For the maximum sentence, the interaction between the report’s conclusion and a priori endorsement of rehabilitation and deterrence was also not significant (ps > .240), but we found a significant interaction between the report’s conclusion and a priori endorsement of retribution, β = 0.40, t = 2.62, p = .010, 95% CI [0.09, 0.71]. The conditional effects showed that participants who strongly endorsed the retribution goal (+1 SD) chose longer maximum sentences in the high recidivism risk condition, β = 1.08, t = 2.24, p = .027, 95% CI [0.12, 2.04]. This effect was not significant for participants who moderately or weakly endorsed retribution (β = 0.18, t = 0.53, p = .596, 95% CI [−0.49, 0.84] β = −0.72, t = −1.51, p = .135, 95% CI [−1.67, 0.23], respectively).

6

We cannot exclude the possibility that additional information suggesting the offender's difficulties in controlling his or her emotions (in the high recidivism risk condition) may also have played a mitigating role in sentencing. Nevertheless, since the conclusions of a clinical report are generally detailed, this information was, in our opinion, necessary to ensure the credibility of the expertise.

Supplemental data

Supplemental data for this article can be accessed here.

References

  1. Arce, R., Farina, F., & Egido, A. (1996). Dynamiques dans la décision des jurés [Dynamics in jury decision-making]. Connexions, 67, 111–120. [Google Scholar]
  2. Atanasova-Denié, Z., & Tostain, M. (2008). Les processus d’attribution de punitions: Étude des relations entre gravité de l’infraction pénale, caractéristiques de l’auteur, émotions et motivations à punir [Punishment’s attribution processes: Study of the relations between seriousness of the penal infringement, characteristics of the author, emotions and motives for punishment]. Les Cahiers Internationaux de Psychologie Sociale, Numéro 78(2), 21–34. doi: 10.3917/cips.078.0021 [DOI] [Google Scholar]
  3. Beccaria, C. (1764). Des delits et des peines [On crimes and punishments] (M. Chevallier, Trans.). Garnier-Flammarion. [Google Scholar]
  4. Bentham, J. (1802). Traités de législation civile et pénale [Treaties on civil and criminal legislation]. Bossange. [Google Scholar]
  5. Bordel, S., Vernier, C., Dumas, R., Guingouain, G., & Somat, A. (2004). L’expertise psychologique, élément de preuve du jugement judiciaire? [Psychological report, element of evidence of judicial judgment]. Psychologie Française, 49(4), 389–408. doi: 10.1016/j.psfr.2004.07.004 [DOI] [Google Scholar]
  6. Carlsmith, K. M. (2008). On justifying punishment: The discrepancy between words and actions. Social Justice Research, 21(2), 119–137. doi: 10.1007/s11211-008-0068-x [DOI] [Google Scholar]
  7. Carlsmith, K. M., & Darley, J. M. (2008). Psychological aspects of retributive justice. Advances in Experimental Social Psychology, 40, 193–236. doi: 10.1016/S0065-2601(07)00004-4 [DOI] [Google Scholar]
  8. Carlsmith, K. M., Darley, J. M., & Robinson, P. H. (2002). Why do we punish? Deterrence and just deserts as motives for punishment. Journal of Personality and Social Psychology, 83(2), 284–299. doi: 10.1037//0022-3514.83.2.284 [DOI] [PubMed] [Google Scholar]
  9. Carlsmith, K. M., Monahan, J., & Evans, A. (2007). The function of punishment in the "civil" commitment of sexually violent predators. Behavioral Sciences & the Law, 25(4), 437–448. doi: 10.1002/bsl.761 [DOI] [PubMed] [Google Scholar]
  10. Cusson, M. (1987). Pourquoi punir? [Why punish?]. Dalloz. [Google Scholar]
  11. Darley, J. M., Carlsmith, K. M., & Robinson, P. H. (2000). Incapacitation and just deserts as motives for punishment. Law and Human Behavior, 24(6), 659–683. doi: 10.1023/A:1005552203727s [DOI] [PubMed] [Google Scholar]
  12. Desmarais, S. L., Johnson, K. L., & Singh, J. P. (2016). Performance of recidivism risk assessment instruments in US correctional settings. Psychological Services, 13(3), 206–222. doi: 10.1037/ser0000075 [DOI] [PubMed] [Google Scholar]
  13. Dumas, R., & Esnard, C. (2019). Confirmatory information processing in legal decision: Effect of intimate conviction. Journal of Police and Criminal Psychology, 34(3), 246–253. doi: 10.1007/s11896-019-9316-2 [DOI] [Google Scholar]
  14. Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. The American Psychologist, 49(8), 709–724. doi: 10.1037//0003-066x.49.8.709 [DOI] [PubMed] [Google Scholar]
  15. Esnard, C., & Dumas, R. (2019). L’intime conviction: entre cadres légaux, représentations et pratiques chez les magistrats et jurés en cour d’assises [The intimate conviction: between legal frameworks, representations and practices among magistrates and jurors in the Assize Courts. Bulletin de Psychologie, Numéro559(1), 53–69. doi: 10.3917/bupsy.559.0053 [DOI] [Google Scholar]
  16. Esnard, C., Dumas, R., & Bordel, S. (2013). Effects of the instruction of ‘intime conviction’ on judicial information processing. European Review of Applied Psychology, 63(2), 121–128. doi: 10.1016/j.erap.2012.12.002 [DOI] [Google Scholar]
  17. Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. doi: 10.3758/BF03193146 [DOI] [PubMed] [Google Scholar]
  18. Feeley, M., & Simon, J. (1992). The new penology: Notes on the emerging strategy of corrections and its implications. Criminology, 30(4), 449–474. doi: 10.1111/j.1745-9125.1992.tb01112.x [DOI] [Google Scholar]
  19. Garapon, A. (1997). Bien juger. Essai sur le rituel judiciaire [To judge well. Essay on judicial ritual]. Odile Jacob. [Google Scholar]
  20. Graham, S., Weiner, B., & Zucker, G. S. (1997). An attributional analysis of punishment goals and public reactions to O. Personality and Social Psychology Bulletin, 23(4), 331–346. doi: 10.1177/0146167297234001 [DOI] [Google Scholar]
  21. Hamilton, V. L., & Rytina, S. (1980). Social consensus on norms of justice: Should the punishment fit the crime? American Journal of Sociology, 85(5), 1117–1144. doi: 10.1086/227127 [DOI] [Google Scholar]
  22. Harcourt, B. E. (2011). Surveiller et punir à l’âge actuariel: Généalogie et critique [Challenging the actuarial method in penal systems: Genesis and critique]. Déviance et Société, 35(1), 5–33. doi: 10.3917/ds.351.0005 [DOI] [Google Scholar]
  23. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. The Guilford Press. [Google Scholar]
  24. Howard, G. S. (1980). Response-shift bias: A problem in evaluating interventions with pre/post self-reports. Evaluation Review, 4(1), 93–106. doi: 10.1177/0193841X8000400105 [DOI] [Google Scholar]
  25. Jacoby, J. E., & Cullen, F. T. (1998). The structure of punishment norms: Applying the Rossi-Berk model. The Journal of Criminal Law and Criminology (1973-), 89(1), 245–312. doi: 10.2307/1144222 [DOI] [Google Scholar]
  26. Kant, E. (1980). Critique de la raison pure [Critique of pure reason] (A. J. -L. Delamarre, & F. Marty, Trans.). Gallimard. (Original work published 1781). [Google Scholar]
  27. Krauss, D. A., & Lee, D. H. (2003). Deliberating on dangerousness and death: Jurors’ability to differenciate between expert actuarial and clinical predictions of dangerousness. International Journal of Law and Psychiatry, 26(2), 113–137. doi: 10.1016/S0160-2527(02)00211-X [DOI] [PubMed] [Google Scholar]
  28. Krauss, D. A., & Sales, B. D. (2001). The effects of clinical and scientific expert testimony on juror decision making in capital sentencing. Psychology, Public Policy, and Law, 7(2), 267–310. doi: 10.1037/1076-8971.7.2.267 [DOI] [Google Scholar]
  29. Krauss, D. A., Lieberman, J. D., & Olson, J. (2004). The effects of rational and experiential information processing of expert testimony in death penalty cases. Behavioral Sciences & the Law, 22(6), 801–822. doi: 10.1002/bsl.621 [DOI] [PubMed] [Google Scholar]
  30. Krauss, D. A., McCabe, J. G., & Lieberman, J. D. (2012). Dangerously misunderstood: Representative jurors' reactions to expert testimony on future dangerousness in a sexually violent predator trial. Psychology, Public Policy, and Law, 18(1), 18–49. doi: 10.1023/A:1022326807441 [DOI] [Google Scholar]
  31. Krauss, D., & Scurich, N. (2014). The impact of case factors on jurors’ decisions in a sexual violent predator hearing. Psychology, Public, Policy, and Law, 20(2), 135–145. doi: 10.1037/law0000007 [DOI] [Google Scholar]
  32. Marsden, E., & Torgerson, C. J. (2012). Single group, pre-and post-test research designs: Some methodological concerns. Oxford Review of Education, 38(5), 583–616. doi: 10.1080/03054985.2012.731208 [DOI] [Google Scholar]
  33. Mary, P. (2001). Pénalité et gestion des risques: Vers une justice ‘actuarielle’ en Europe? [Penalty and risk management: Toward actuarial justice in Europe?]. Déviance et Société, 25(1), 33–51. doi: 10.3917/ds.251.0033 [DOI] [Google Scholar]
  34. McFatter, R. M. (1978). Sentencing strategies and justice: Effects of punishment philosophy on sentencing decisions. Journal of Personality and Social Psychology, 36(12), 1490–1500. doi: 10.1037/0022-3514.36.12.1490 [DOI] [Google Scholar]
  35. Monnery, B. (2016). Prison, reentry and recidivism: Micro-econometric applications. Economies and Finances [Doctoral dissertation]. https://tel.archives-ouvertes.fr/tel-01452006/document
  36. Monohan, J., & Steadman, H. (1994). Toward the rejuvenation of risk research. In Monohan J. & Steadman H. (Eds.), Violence and mental disorder: Developments in risk assessment (pp. 1–17). University of Chicago Press. [Google Scholar]
  37. Moulin, V., & Palaric, R. (2013). À propos de quelques fonctions des expertises judiciaires au pénal [Concerning certain functions of legal expertise in criminal justice]. L'information Psychiatrique, 89(9), 713–721. doi: 10.3917/inpsy.8909.0713 [DOI] [Google Scholar]
  38. Niang, A., Testé, B., & Leclerc, C. (2019). Analyse psychosociale du rôle des motivations à punir dans l’usage de l’expertise d’un accusé en contexte de détermination de la peine en France [Psychosocial analysis of the role of goals of punishment in the use of a defendant's expertise in the context of sentencing in France] [Doctoral dissertation]. Université Rennes 2, France. [Google Scholar]
  39. Orth, U. (2003). Punishment goals of crime victims. Law and Human Behavior, 27(2), 173–186. doi: 10.1023/A:1022547213760 [DOI] [PubMed] [Google Scholar]
  40. Oswald, M. E., Hupfeld, J., Klug, S. C., & Gabriel, U. (2002). Lay-perspectives on criminal deviance, goals of punishment, and punitivity. Social Justice Research, 15(2), 85–98. doi: 10.1023/A:1019928721720 [DOI] [Google Scholar]
  41. Oytana, Y. (2016). Les coûts des expertises judiciaires dans les procédures inquisitoire et accusatoire [The cost of judicial expertise in the inquisitorial and adversarial procedures]. Revue Économique, 67(1), 5–48. doi: 10.3917/reco.pr2.0050 [DOI] [Google Scholar]
  42. Reifman, A., & Keyton, K. (2010). Winsorize. In Salkind N. J. (Ed.), Encyclopedia of research design (pp. 1636–1638). SAGE. [Google Scholar]
  43. Robinson, E. A., & Doueck, H. J. (1994). Implications of the pre/post/then design for evaluating social group work. Research on Social Work Practice, 4(2), 224–239. doi: 10.1177/104973159400400207 [DOI] [Google Scholar]
  44. Shuman, D. W., & Sales, B. D. (1998). The admissibility of expert testimony based upon clinical judgment and scientific research. Psychology, Public Policy, and Law, 4(4), 1226–1252. doi: 10.1037/1076-8971.4.4.1226 [DOI] [Google Scholar]
  45. Tostain, M. (2007). L’influence des motivations à punir sur les jugements de responsabilité et l’attribution de sanction aux auteurs d’infractions pénales [The influence of the motives for punishment on judgments of responsibility and the attribution of sanctions to the perpetrators of criminal offences]. Les Cahiers Internationaux de Psychologie Sociale, Numéro 75–76(3), 35–49. doi: 10.3917/cips.075.0035 [DOI] [Google Scholar]
  46. Tostain, M., & Lebreuilly, R. (2013). L'influence des logiques pénales rétributives et utilitaristes sur le niveau d'acceptation de la rétention de sûreté dans les cas de crimes pédophiles [Influence of retributive and utilitarian factors on the acceptance of civil commitment for pedophiles]. Les Cahiers Internationaux de Psychologie Sociale, Numéro 98(2), 237–257. doi: 10.3917/cips.098.0237 [DOI] [Google Scholar]
  47. Van Wingerden, S., Van Wilsem, J., & Moerings, M. (2014). Pre-sentence reports and punishment: A quasi-experiment assessing the effects of risk-based pre-sentence reports on sentencing. European Journal of Criminology, 11(6), 723–744. doi: 10.1177/1477370814525937 [DOI] [Google Scholar]
  48. Vidmar, N., & Miller, D. (1980). Social psychological processes underlying attitudes toward legal punishment. Law & Society Review, 14(3), 565–602. doi: 10.2307/3053193 [DOI] [Google Scholar]
  49. Viljoen, J. L., Cochrane, D. M., & Jonnson, M. R. (2018). Do risk assessment tools help manage and reduce risk of violence and reoffending? A systematic review. Law and Human Behavior, 42(3), 181–214. doi: 10.1037/lhb0000280 [DOI] [PubMed] [Google Scholar]
  50. Warner, K., Davis, J., Spiranovic, C., Cockburn, H., & Freiberg, A. (2019). Why sentence? Comparing the views of jurors, judges and the legislature on the purposes of sentencing in Victoria. Criminology & Criminal Justice, 19(1), 26–44. doi: 10.1177/1748895817738557 [DOI] [Google Scholar]
  51. Weiner, B., Graham, S., & Reyna, C. (1997). An attributional examination of retributive versus utilitarian philosophies of punishment. Social Justice Research, 10(4), 431–451. doi: 10.1007/BF02683293 [DOI] [Google Scholar]

Articles from Psychiatry, Psychology, and Law are provided here courtesy of Taylor & Francis

RESOURCES